Missing Unicode license

[stefan6419846]

The idna package currently ships data derived from the Unicode data files according to the generator script at https://github.com/kjd/idna/blob/master/tools/idna-data The corresponding README file states that the files are subject to the Unicode-3.0 license, which requires shipping the notice and I could not find in this repository: https://www.unicode.org/Public/17.0.0/ReadMe.txt https://www.unicode.org/license.txt

[kjd]

While I am not opposed to including the licensing agreement, nothing from the standard is directly included in the repository, so I am not sure that is appropriate. As you say there is data that is derived and transformed from analyzing Unicode properties, which is factual data and facts are not generally copyrightable.

For comparison I looked at what cpython does since it more directly includes Unicode data in the unicodedata module. It does not appear to include the Unicode license for that derivation either. See https://github.com/python/cpython/blob/main/Tools/unicode/makeunicodedata.py

[stefan6419846]

Thanks for the fast response. While facts might not be copyrightable, the specific dataset can still be considered copyrightable (for example due to database laws).

I am aware that idna clearly is not the only library like this and that violations are unlikely to be sanctioned, but as long as including the notice does not cause large trouble, it simplifies license compliance handling and contributes to a FOSS ecoystem with proper attribution chains.

[StanFromIreland]

While I am not opposed to including the licensing agreement, nothing from the standard is directly included in the repository, so I am not sure that is appropriate. As you say there is data that is derived and transformed from analyzing Unicode properties, which is factual data and facts are not generally copyrightable.

I agree, I apply similar logic to unicodedata in python/cpython#154493 (comment).